Systems are not static. They have to be adapted to changing market conditions an need to be regularly updated time to time. All forced (external) chances make traders (even with high tech equipment and recently great signals) to loose money or reduce their chances to perform best. To get successful or stay successful - traders need innovation. Innovation means changing or developing systems for current market conditions and 'testing' them with many real trades in the future. Traders take risk in performing real trades because they are confident to run dozens of money-making trades before changing their system. All the money can be lost worst case during that time. The most overlooked aspect is money management which is a function of risk aversion and of edge. In addition a system must suit the traders personality. ------------------------------------------------------------------------------- Pupose of this thread is to openly share my ideas and systems for the FDAX and to get feedback from enthusiastic FDAX traders. ------------------------------------------------------------------------------- I very much welcome posts of peer traders who show how they worked hard to find the currently best FDAX systems for them. Its also interesting to see what time frame they prefer and how they manage current market conditions. Ideally all the readers and posters of this new thread should profit from the ideas and get educational knowledge to create their own systems or to merge existing systems into a new or compounded one.
Possible FDAX counter wave system is to go against the waves at 20 points and a profit target of 12.5 points as shown below. Stops and money management need to be applied. What do you think? Please post your ideas!
i also trade the dax and to the best of my knowledge, the dax is better read in 1H aggregation. i work with smaller aggregations including tick but the coherence of the signals never match the ones i get in 1H. re the signals itself, i am pretty comfortable with my prop system based on support and resistance. there are a couple of aspects that i had to work out in order to get an acceptable performance though. 1. i use an algo to analyze the present vola in the dax and derive the future vola. then i use this output to setup the projected s&r band. i have spent much time in this in order to come up with a bot that adapts itself to constant changing vola conditions. imo, any static system cannot work effectively throughout the course of any (acceptable) time frame considered. 2. in addition i use a macro trend indicator that prevents entering trades in counter trend. 3. i dont use SL or PT because in the long run they proved to cap the profit potential and increase comms and slippage wo any visible advantage.
asap I think 60 minutes are a good timeframe! I am very interested in VDAX and VIX and I am playing around with some spreadsheets. I my experience the second and third derviative have negative correlation on a daily base at least for the INDU. I didn't perform the data for the DAX yet. It would be also interesting to see an autocorrelation on the 60 min and 15 min time base. I am not using any indicator - it's pure FDAX. In addition I look at all major indexes to get a feeling for the market direction and to search some correlations. The DAX Index is also a great tool. If a system on the DAX is more profitable than a system on the FDAX, I will use the DAX as signal provider for the FDAX trades. I can also use Digital options and other vehicles to trade the DAX-Future. In addition there is a big option market at the EUREX.
Potential FDAX system based on 2.5 point grid: 1) go with each wave at 3 grid units (7.5 points + 1 tick) using stop orders 2) set stops at -3 grid units (7.5 points + 1 tick) trail stop on grid 3) take profit at 6 grid units (15 points) to make a profit of 3 grid units (7.5 points - 1 tick) One point has 5 ticks (5*0.5 = 2.5)
Possible FDAX system based on the 5 point (10 ticks) grid 1) stop in at 10 points + 1 tick This means that the FDAX needs to pass 2 grid units (a 5 points)before I enter. In total min. 12 ticks (2*5+1+1) or max. 16 ticks (2*5+1+5) are needed to trigger the trade. 2) set trailing stop 10 points + 1 tick below (trail on grid) This is the reversed action to protect too big losses. If there is regular trading without gaps and no slippage the stop will cut a potential loss to 12 ticks initially. As soon the FDAX proceeds 1 unit (=5 points or 10 ticks) in the right direction, the stop will be trailed on the grid. That mean that the stop is always one tick lower (or higher for shorts) than the grid line. 3) take profit at 20 points That means get out when the profit target is reached. No emotions, no compromise. Profit is 2 units - 1 tick (9 points or 18 ticks) ----------------------------------------- I call that system FDAX5_2 S242 ----------------------------------------- The system can be improved by taking trades only in the direction of the trend. The system has been developed by taking the most recent FDAX data (only a few hours/days). Is that helpful for you? Please post your experience in developing the currently best FDAX system for you.
good job gg. i like the dynamic bands as said earlier. for instance, given a certain weekly trend in the dax, one is able to determine a variance band that is projected into the future for probability analisys. this example shows the band for the previous week. as you can see, the dax was stuck in the band during all week, even though the vola went down a bit later, the band set the stage for hi prob entry and exit setups across the week. this is a 4M agg chart, obviously a 1H chart gives a much clearer picture and keeps its coherence longer, while the 4M typically needs a band adjustment per week. either way, the 4M is a good tool to make quick intraday trades. one other interesting aspect of the dax bands is the incredible correlation between es and the dax. this allows to base es trades on dax band as well and most times they turn out right whenever the dax band is not violated. the way this is traded is when the price gets closer to one of the bands and confirms a bounce. the protective SL should be set 5 points bellow the band and PT should be set 5-10 points before the opposite band. the average point gain is 40 (approx half of the band) while the average loss is 20 or so points.
Hi gg12, In your experience, what kind of systems do you find more consistent and profitable? is it those which go with the trend like entering after 10 points and going for another 10 points or those that fade the move like fading a 20 point move going for 12.5 points in your previous post? And also it helps if you put in the green and red arrows to show your entries and exits. I find that there is certainly some "curve-fitting" involved here where you look at some past data and come out with a set of values that is profitable. And do you always work only with 15 min chart and why? Thanks
tyrant You are right. I am curve fitting, because of using historical data. On the other hand the systems are so simple and universal that I think they can be used in similar market conditions in the future. According time frames I am using everything, from 1 minute to some years. I aim to overlay the systems in a way that signals x out each other. From my temperament I am a very short time trader. I like immediate results, but I have also learned to be patient - sometimes very patient and persistant. For this thread I will provide the best system available. I will merge all input I will find on the web and from peers. I you have great input I will include it in my system. For the signals. Green means successful trade, red means no successful and orange means scratch trade. A scratch trade after commission is always a loosing trade. Stop entries or exits have the risk of slippage. Overnight positions (and even intraday positions) have the risk of gaps. I am going with and against moves. I am going long and short. I am doing scalps, momentum and swing trades. I am familiar with options as well. I am open for all aspects of trading. Hope that helps. I will provide everything needed.